Rough sets and their variants
Mahmut Dirik
Abstract
Fire is a natural disaster that poses a profound existential threat to humanity. It has traditionally been fought with conventional methods, which, unfortunately, are often fraught with limitations and potential environmental damage. Given these limitations, there is an urgent need for research into ...
Read More
Fire is a natural disaster that poses a profound existential threat to humanity. It has traditionally been fought with conventional methods, which, unfortunately, are often fraught with limitations and potential environmental damage. Given these limitations, there is an urgent need for research into novel firefighting methods. Sound wave-based firefighting systems, an emerging solution, show promising potential in this regard.The current study uses an extensive data set derived from numerous experimental trials of sound-wave-based firefighting. Based on this extensive dataset, we have developed a sound wave technology-based fire suppression model that includes five different fuzzy logic methods: Fuzzy Rough Set (FRS), Fuzzy K-Nearest Neighbors (FNN), Fuzzy Ownership K-Nearest Neighbors (FONN), Fuzzy-Rough K-Nearest Neighbors (FRNN), and Vaguely Quantified K-Nearest Neighbors (VQNN).The main objective of these models is to accurately distinguish between the extinguished and non-extinguished states of a flame. This classification is based on a number of intrinsic model parameters, such as the type of fuel, the size of the flame, the decibel level, the frequency, the airflow, and the distance.To evaluate the classification effectiveness of the models, a number of statistical methods were used, including Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Kappa Statistics (KP), and Mean Square Error (MSE).Our analysis yielded promising results, with the models FRS, FNN, FONN, FRNN, and VQNN achieving classification accuracies of 93.12%, 96.66%, 95.56%, 96.35%, and 96.89%, respectively. These results confirm the high accuracy of the proposed model in classifying fire data and underline its practical applicability.
Fuzzy multisets and their variants
Mahmut Dirik
Abstract
Due to developments in printing technology, the number of counterfeit banknotes is increasing every year. Finding an effective method to detect counterfeit banknotes is an important task in business. Finding a reliable method to detect counterfeit banknotes is a crucial challenge in the world of economic ...
Read More
Due to developments in printing technology, the number of counterfeit banknotes is increasing every year. Finding an effective method to detect counterfeit banknotes is an important task in business. Finding a reliable method to detect counterfeit banknotes is a crucial challenge in the world of economic transactions. Due to technological development, counterfeit banknotes may pass through the counterfeit banknote detection system based on physical and chemical properties undetected. In this study, an intelligent counterfeit banknote detection system based on a Genetic Fuzzy System (GFS) is proposed to detect counterfeit banknotes efficiently. GFS is a hybrid system that uses a network architecture to fine-tune the membership functions of a fuzzy inference system. The learning algorithms Fuzzy Classification, Genetic Fuzzy Classification, ANFIS Classification, and Genetic ANFIS Classification were applied to the dataset in the UCI machine learning repository to detect the authenticity of banknotes. The developed model was evaluated based on Accuracy (ACC), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Error Mean, Error STD, and confusion matrix. The experimental results and statistical analysis showed that the classification performance of the proposed model was evaluated as follows: Fuzzy = 97.64%, GA_Fuzzy = 98.60%, ANFIS = 80.83%, GA_ANFIS = 97.72% accuracy (ACC). This shows the significant potential of the proposed GFS models for fraud detection.
Type-2 fuzzy sets and their variants
Mahmut Dirik
Abstract
In this study, a hybrid model for prediction issues based on IT2FLS and Particle Swarm Optimization (PSO) is proposed. The main contribution of this work is to discover the ideal strategy for creating an optimal value vector to optimize the membership function of the fuzzy controller. It should be emphasized ...
Read More
In this study, a hybrid model for prediction issues based on IT2FLS and Particle Swarm Optimization (PSO) is proposed. The main contribution of this work is to discover the ideal strategy for creating an optimal value vector to optimize the membership function of the fuzzy controller. It should be emphasized that the optimized fuzzy controller is a type-2 interval fuzzy controller, which is better than a type-1 fuzzy controller in handling uncertainty. The limiting membership functions of the type-2 fuzzy set domain is type-1 fuzzy sets, which explains the trace of uncertainty in this situation. The proposed optimization strategy was tested using ECG signal data. The accuracy of the proposed IT2FLS_PSO estimation technique was evaluated using a number of performance metrics (MSE, RMSE, error mean, error STD). RMSE and MSE with IT2FI were calculated as 0.1183 and 0.0535, respectively. With IT2FISPSO, these values were calculated as 0.0140 and 0.0029, respectively. The proposed PSO-optimized IT2FIS controller significantly improved its performance under various operating conditions. The simulation results show that PSO is effective in designing optimal type 2 fuzzy controllers. The experimental results show that the proposed optimization strategy significantly improves the prediction accuracy.